As we saw in the previous chapter, all adaptations are for something. The argument in this chapter is that the process of infant–mother attachment is an adaptation for a special kind of learning. I begin by explaining why I approach human development from the perspective of attachment theory. This I do by using the assumption of optimality to identify the jobs that need to be done during development. I then outline attachment theory and research to show that the attachment process does these jobs. Attachment, I shall argue, is for learning about (gaining knowledge about, representing) one's past and one's present in order to predict one's future – and thereby to “evaluate” one's alternatives and “choose” (i.e., not necessarily consciously) one's optimal developmental pathway. Because development itself is ultimately for reproduction, however, I argue as well that the attachment process is ultimately for “choosing” one's optimal reproductive strategy.

In addition to being for something, however, adaptations are relational as well. In the midst of showing how the attachment process is for reproduction, therefore, I will also address the reverse engineering question of how it is relational.